The following column appeared in print on January 21, 2017.
A Glimpse at a New Type of Search Engine
Diane Lynn Tibert
When my sister told me a friend of the family had passed away, I searched for her obituary online. The woman had been born in Newfoundland but had lived many years in the United States and Nova Scotia—where we met her—before she returned to spend her final years in her birth province.
I was unsure which newspaper the obituary would appear in, if it would appear at all. I assumed the funeral home handling the arrangements would post it to their website. I only had to find the right one. A Google search with keywords to locate the obituary didn’t find anything. I began adding and subtracting keywords hoping I’d stumble onto something but my search failed.
The way Google search engine works is pages containing all (or most) of the keywords appear in the results. Quotation marks around certain keywords help narrow the search by finding these words in that particular order, but the results don’t always find the right information.
Internet search engines have improved in the past 25 years but are still flawed. I’ve had much success with them by using their advanced search techniques and by adding things such as quotation marks and the negative sign to eliminate certain words, but I wish things could be easier. I hadn’t given any thought to how they could be improve until I found the genealogy website MooseRoots.
MooseRoots looks like many other genealogy sites. Visitors can enter a name into the search box and find individuals in census, birth, marriage, divorce, death and cemetery records. There’s also a section with immigration records and military records, including those for Canada during the First and Second World Wars.
But what makes MooseRoots different is it uses “Graphiq’s semantic technology to deliver deep insights via data-driven articles, visualizations and research tools”. I’m always researching on the Internet, but I’ve never heard of this type of search engine before. It promises to deliver more accurate results, but how does it work?
Graphiq is the company behind the technology. It began developing the unique search engine in 2010. The answer to how it works is not simple and needs more space than I have here, but the functioning of it can be explained in a simple example Graphiq uses on their website.
Imagine going into a phone store looking for an upgrade from your current device. When the salesperson asks if you are an Apple fan, you immediate know she’s talking about a type of phone, not a fruit. Your brain uses contextual cues to know the type of apple in question to answer correctly. Computers can’t do that…until now.
Graphiq has developed a program that will help their search engine understand the connection between the keywords and their context to produce more accurate results. The semantic search works to choose the correct meaning of the question being asked in the search engine, working with the user’s intent, and filtering out other results that may only contain the keywords in a scrambled mess.
This type of search engine has many possibilities. It not only delivers results in text, it strives to answer questions and provide information using visuals. Graphiq’s “ultimate goal is to build the world’s largest and most interconnected base of human knowledge”. It will be interesting to see how this technology changes our experience on the Internet.